Are Workers’ Remittances Relevant for Credit Rating Agencies? Rolando Avendaño OECD Development Centre Norbert Gaillard Bank for International Settlements Sebastián Nieto Parra OECD Development Centre Abstract Remittances flows are an important source of financing for developing countries. In addition to the microeconomic impact at the household level, remittances have grown into an important pillar of macroeconomic stability, acting to reduce financial vulnerabilities, lessen probability of current account reversals and ultimately reduce credit risk. This paper, assessing the country risk models for the three main Credit Rating Agencies (CRAs), focuses on the role of workers’ remittances in the estimation of the sovereign ratings of 83 emerging countries over four dimensions. First, it explores the role that workers’ remittances play in country risk models. Second, it assesses the extent to which CRAs take remittances into account in determining ratings. Third, it seeks to capture the potential effect that remittances have for sovereign ratings when considered in rating models. Fourth, it attempts to assign sovereign ratings to unrated Latin American countries. We conclude that while CRAs take remittances flows nominally into account as a determinant of ratings, remittances have a real and significant impact on a certain type of sovereign rating, through the reduction of the volatility of external flows as well as the improvement of the solvency ratio. Such effects are concentrated on a set of low and middle income countries, for which the dependence on remittances is high. JEL Classification: O11, F24, F3, G24 Keywords: credit rating agencies, remittances, sovereign ratings, emerging and developing capital markets. THIS DRAFT 09 MAY 20091 I. Introduction 1 The authors are indebted to Sara Bertin, Thomas Dickinson, Kiichiro Fukasaku, Juan de Laiglesia, Javier Santiso and the LEO (Latin American Economic Outlook) team of the OECD Development Centre for helpful comments and discussions. The views expressed in this paper do not necessarily reflect those of the OECD. Mailing address: OECD Development Centre, 2, rue André Pascal, 75775 Paris Cedex 16, France. 1 Workers’ remittances can have an effect in the way that country risk is assessed. Previous research on the access of sovereigns to international capital markets (see Ratha (2006) and World Bank (2006)) suggests that sovereign creditworthiness could be improved by including remittances flows in key indebtedness indicators, such as debt-to-exports and debt service to current account ratios. These have been identified in the literature as common determinants of country ratings. It is worth noting two surveys at the crossroads of the literature on sovereign ratings and remittances flows. First, Ratha et al. (2007) define a standard ratings model and find that a number of unrated countries would be likely to have higher ratings than expected, notably on account of foreign currency inflows such as remittances. According to Ratha (2006), “country credit ratings by major international rating agencies often fail to account for remittances”. Second, rating agencies note in their country studies that remittances matter to determine ratings for countries in which this flow is considerable. Fitch (2008) underlines that remittance flows may positively impact ratings (e.g. El Salvador)2. In its outlook for Mexico, Standard and Poor’s (2005) stresses remittances’ importance as an income source for the balance of payments, and their impact on other determinants of sovereign ratings, such as public finances. More recently, Moody’s highlights that, for a country like the Philippines, a slower economic growth for 2009 would also be explained by a decline in remittances, which account for more than 10% of domestic output and are a major driver of consumption3. Despite these stylized facts, little research has been devoted to analyse the impact that remittances have on sovereign ratings assigned by CRAs. Our paper attempts to address this question by building a rating model over a long time span (1993-2006), and estimating ratings for a sample of 83 emerging countries and the three main CRAs. Concretely, this paper seeks to answer three key questions: Do rating agencies really take remittances fully into account in their analyses? How can we capture the effect of remittances on ratings? And finally, what is the potential effect of remittances when included in market variable estimations? With this purpose, it is crucial to understand why CRAs should take remittances into consideration when assigning ratings. Although the effects of workers’ remittances at the macro and the micro level have been largely studied (see World Bank, 2006 for a review of the literature), the evidence on the implications for ratings is still scarce (World Bank 2006, Ratha 2007). The question is crucial given the importance of remittances flows towards the developing world, with Latin America being no exception. Over the last twenty years remittances have increased considerably, even by comparison to other capital flows to developing economies. Currently, remittances account for around one third of total financial flows to the developing world. With foreign direct investment (FDI), remittance flows to developing countries overtook Official Development Assistance (ODA) in the nineties and for some developing countries reached comparable and even higher levels than FDI flows. In that context, Ratha (2004) and Ratha, Mohapatra and Xu (2008) provide an in-depth analysis of the importance of remittances since the late 1990s. In Latin American and Caribbean countries the increase in remittances flows is also considerable and their proportion is high with respect to other external flows. During the period 1970-2006, remittances accounted for about 50 percent of FDI and were twice as large as ODA flows. In 2006, remittances 2 3 See for instance Standard and Poor's (2005) and Fitch (2008). “Moody's: Slowing remittances hurt RP”, Manila Bulletin, February 14, 2009 (online article). represented more than 80 percent of FDI and more than eight times ODA flows. This is particularly clear in some Central American countries. During the same year, the volume of remittances towards developing countries reached about 70 percent and 85 percent of FDI and ODA flows respectively. Figure 1. Net flows to developing countries 300 250 USD Billions 200 150 100 50 -50 2006 2004 2002 2000 1998 1996 1994 1992 1990 1988 1986 1984 1982 1980 1978 1976 1974 1972 1970 1968 1966 1964 1962 1960 0 Year ODA FDI Inflows Bonds Equity Remittances Source: Authors, based on World Bank and OECD data, 2009. At the macro-economic level, an important characteristic concerns remittances’ impact on the balance of payments, in particular when compared with other capital flows. Figure 2 exhibits the volatility of the major capital flows to emerging countries over the period 1990-2006. Figure 2. Volatility of flows by period 1990-2006 Volatility of Inward Flows 0.018 1990-1999 0.016 2000-2007 0.014 0.012 0.01 0.008 0.006 0.004 0.002 0 Remittances Portfolio FDI Source: Authors based on World Bank and OECD data 3 ODA Exports Note: the volatility of a quantity is defined as its coefficient of variation (the standard deviation of the quantity divided by the mean of its absolute value). The normalisation avoids finding larger volatilities for larger flows. It is calculated for each recipient and then averaged over all developing countries in the sample. By comparing the volatility of remittances with respect to other flows we notice they have a lesser volatility with respect to portfolio flows (equity and bond flows) but also with respect to FDI flows. Even regarding ODA flows, remittances have a lower volatility than other capital flows. This effect is clearer in the later years of the sample. Importantly, the correlation between remittances and other flows is small, contributing to the stability of the total external flows. The correlation of remittances with FDI and ODA are close to 0.20 and 0.0 respectively. Migrants’ remittances constitute a large source of foreign capital in many Latin American countries and can be considered as a stable source of financing compared with other financial flows. Remittances, in the same way as foreign investment or exports, are important items in the balance of payments, contributing to mitigate credit risk at the country level. More precisely, remittances strengthen financial stability by reducing the probability of current account reversals (Bugamelli and Paterno, 2005). This, in turn, can be related to the probability of default studied in country risk models. In the same way, remittances can have a countercyclical effect in most countries of the region, thus significantly reducing growth volatility (Fajnzylber and Lopez 2007). The overall conclusion of Close to Home, the comprehensive World Bank study on Latin America, is that remittances are an engine for development, but that they are neither “manna from heaven” nor a substitute for sound development policies. Empirical evidence shows that remittances can have a positive impact on economic growth and poverty reduction in developing countries by allowing capital accumulation at the household level (Giuliano and Marta Ruiz-Arranz , 2005; Osili, 2006; Dustmann and Kirchamp (2001). For the case of Latin American countries, we also find a positive effect of remittances at the microeconomic level (see Massey and Parrado , 1998 for the case of Mexico and Cardona Sosa and Medina, 2006 for Colombia). However, as pointed out elsewhere, migrant-based income can become costly to emerging countries when resources are mismanaged. Remittances may reduce the government’s incentive to maintain fiscal policy discipline (Chami et al., 2008). Moreover, this dependence raises a moral hazard problem by reducing the political will to implement reforms and pushing real exchange appreciation. These findings are consistent with Amuedo-Dorantes and Pozo (2004) who relate higher remittances flows to the reduction of the receiving country’s competitiveness. Remittances should matter when country risk is assessed in developing countries given the direct effect that they have on the balance of payments (e.g., stability in the capital flows, exchange rates) and more generally on the real economy (e.g., economic growth, output volatility, poverty reduction). Indeed, remittances can play a positive role in the credit risk of a country and this impact can be observed directly (e.g., stability of balance of payments) or indirectly (e.g. reduction of income inequality, creation of new firms). For Latin American countries, the central empirical analysis of this paper, remittances are crucial, given their high level with respect to the size of the economies (e.g., Guyana, Honduras, Haiti, Jamaica, El Salvador) or given their high level in absolute values (e.g., Mexico, Brazil, Colombia, Guatemala, El Salvador, Ecuador). Differences among countries in the region remain important. Figure 3 shows that remittances are particularly relevant for Central American and Caribbean countries. Figure 3. Remittances in Latin America – 2006 25% 25000 Workers Remittances Chile Venezuela, RB Brazil Argentina Uruguay Panama Dominica Peru Colombia Costa Rica Mexico Bolivia Paraguay 0 Belize 0% Ecuador 5000 Dominican … 5% Guatemala 10000 Nicaragua 10% Jamaica 15000 El Salvador 15% Haiti 20000 Guyana 20% Honduras % 30000 Remittances/GDP USD Millions 30% Source: Authors calculation, based on Global Development Finance, 2009. Workers’ remittances can therefore have both direct and indirect impacts on national economies. In this paper we analyse the relationship between remittances and country credit risk through the balance of payment channel. More precisely, we study the impact of remittances on country’s creditworthiness via two dimensions. First, we analyse a common channel to measure remittances on country risk (Ratha 2005, World Bank 2006). By including remittances in a traditional solvency ratio studied in the ratings literature (i.e., the ratio of debt to exports of goods and services), remittances can improve sovereign ratings. Second, we introduce volatility of external flows as an additional variable explaining sovereign ratings. This variable is related to the variability of trend of major inward external flows (FDI flows, Portfolio flows, ODA, Bank loans, Exports and Remittances.). These flows are particularly important for Latin America, where saving rates are low and dependence on external financing high. Our results show that remittances can reduce volatility of external flows given their stability (when compared to other flows) and low correlation with other external flows. The remainder of this article is organised as follows. In section II, we provide a review of the literature on sovereign ratings and in particular on the relevance of sovereign ratings for emerging economies as well as the determinants of these ratings. Section III presents the most important stylised facts and analyses the results of the econometric model. In particular, this section analyses the impact of remittances flows on four representative models of the literature in ratings and our own model, where a counterfactual analysis is presented. We also provide an empirical analysis for countries with a high share of remittances (as a percentage of GDP). Finally, section IV provides concluding remarks and sketches the major policy implications that follow from this research. 5 2. Review of the literature In this section we provide first some empirical evidence on the relevance of CRAs for capital markets and more especially for emerging economies. Second, we present the main results of the seminal papers related to the determinants of ratings. It has been pointed out that “the recent financial market turbulence has brought credit rating agencies under fire” and academia as well as policy-makers argue for a reform of the business model of CRAs (Portes 2008). Rating agencies are faced with a serious conflict of interest, to the extent that their remuneration comes from rated issuers (for a theoretical analysis see Mathis, McAndrews and Rochet, 2008), both in the context of public of private borrowers. This is a crucial issue, given CRAs’ considerable and increasing role on international capital markets. In this context, there is a large and useful literature studying the impact of ratings on market prices and bond spreads. Focusing on market prices, Kaminsky and Schmukler (2001) find that downgrades and upgrades have an impact on country risk and stock returns: these rating changes are transmitted across countries, with neighbour-country effects being more significant. They conclude that rating agencies may contribute to heighten financial instability. The study of sovereign risk assessment has focused on comparing ratings to market spreads. For the period 1987-1994, Cantor and Packer (1996) find a greater impact on spreads from a rating change in the case of Moody’s or if it is related to speculative-grade countries. Reisen and Von Maltzan (1999) show that, during the period 1989-1997, Fitch, Moody’s and S&P’s downgrades have a significant impact on spreads, contrary to upgrades, which were anticipated by the market. For them, sovereign ratings have the potential to moderate euphoria among investors on emerging markets but rating agencies failed to exploit that potential in the 1990s. Sy (2001) highlights the strong negative relationship between ratings and EMBI+ spreads declines during periods of high risk aversion (e.g., 1997-1998). Mora (2006) examines Moody’s and S&P’s ratings and concludes that the procyclicality of ratings is not ascertained when considering the post Asian crisis years. Analyzing sovereign ratings issued by the three agencies for 19932007, Gaillard (2009) finds that the procyclicality of ratings was much sharper during periods of high risk aversion (1997-1998 in particular) than periods of low risk aversion (2005-2007). He also highlights the greater stability of Moody’s ratings. In a different way, Cavallo et al (2008) develop a simple Hausman specification test and find that there is some informational content in sovereign ratings that is not completely captured by market spreads. Additional tests reinforce their conclusion that ratings matter. Lastly, going beyond the traditional “ratings vs. spreads” view, Roubini and Manasse (2005) present an original sovereign risk assessment methodology by using a binary recursive tree. This enables them to better discuss appropriate policy options to prevent crises. The literature focusing on sovereign ratings methodology has expanded since the mid 1990s. Cantor and Packer (1996) identify five variables that may explain S&P and Moody’s sovereign ratings: per capita income, inflation, external debt, level of economic development and default history. Jüttner and McCarthy (2000) show that Cantor and Packer’s model becomes less accurate after the Asian crisis. They suggest that the determinants of 1998 ratings are the current account balance, the indicators for economic development and default history, the interest rate differential vis-à-vis the USD, and the range of problematic assets. Nevertheless, several follow-up studies corroborate Cantor and Packer’s results. For Afonso (2003), most significant variables for 2002 ratings (per capita income, inflation, indicators for economic development and default history) are already determinants for Cantor and Packer. Moody’s own study (Moody’s 2004) produces a similar finding: two of their four explanatory variables (per capita GDP and external debt) are the same as Cantor and Packer’s. The main finding of Moody’s is the incorporation of a political variable that significantly improves the model. For Rowland (2005), the level of international reserves, as a share of GDP, and the openness of the economy are additional relevant determinants. Sutton (2005)’s findings are consistent with previous papers. He also argues that the maturity structure of structure of international banking claims against both private and public sector entities in the country is a significant variable. 7 3. Empirical model 3.1.1. Data Description As noted in the previous section, the literature on sovereign ratings is extensive. We have tried to focus on the most representative work to identify the variables considered by agencies when assigning a rating to public borrowers. The traditional approach in the literature has been to regress the dependent variable (sovereign rating) on a series of macroeconomic and institutional indicators. Table 1 summarizes the period and variables used by Cantor and Packer (1996), Rowland and Torres (2004), Sutton (2005) and Mora (2006) to analyze the determinants of sovereign ratings. Whereas Cantor and Packer and Sutton base their analysis on a cross-country study, Rowland and Torres and Mora use panel data to estimate rating determinants. Most of these studies use one or more of the available ratings published by the three main rating agencies, Standard and Poor’s, Moody’s and Fitch. Table 1 contains also our preferred model, that summarizes our analysis of previous rating models. Table 1. Summary of models and variables • • • • • • • • Spread (lag) • • • • • Dummy default on bonds and bak debt • • Dummy European Union • • • Corrruption Index • • • Indicator of Economic Development 1993-2008 • Default variable (diff. for each model) 2009 • Default History Our model • EMBI Dynamic • Volatility External Flows • Ratios short-term bank/Total claims 1986-2001 Openness ((Exports + Imports)/GDP) 2006 • • • Current Account Balance/GDP Mora • Reserves/GDP • Reserves 2004 Spread 1987-2001 2005 Debt Service/GDP • 2004 Sutton • Total debt/ Exports • Rowland and Torres • External debt/exports • • External Debt / Exports Fiscal balance • 1995 External Debt/GDP Inflation • Period 1996 External balance GDP growth • Year Cantor and Packer Fiscal Balance/GDP Income per capita Independent Variables Sovereign Rating Institutional rating Dependent Variable • • • • Source: the authors based on Cantor and Packer (1996), Rowland and Torres (2004), Sutton (2005) and Mora (2006) The results presented in the table above are straightforward. Sovereign ratings are associated to a country’s fundamentals and, in contrast with sovereign spreads (e.g., Eichengreen and Mody, 2006), endogenous factors are only analysed. More precisely, macroeconomic conditions (e.g., inflation rate, GDP growth), solvency ratios (e.g., external debt over exports, external debt service over GDP) and structural aspects (e.g., GDP per capita, economic development) are employed as determinants of sovereign ratings. In this paper we use data on annual ratings from the three main rating agencies: Standard and Poor’s, Moody’s and Fitch. The covered period is 1993-2006, the frequency is annual and the initial sample includes 83 rated countries (excluding High Income countries according to World Bank’s definition). Data on ratings has been transformed linearly according to Table 2. Table 2. Rating Linear Transformation S&P's Fitch Moody's Rating AAA AA+ AA AAA+ A ABBB+ BBB BBBBB+ BB BBB+ B BCCC+ CCC CCCCC C SD AAA AA+ AA AAA+ A ABBB+ BBB BBBBB+ BB BBB+ B BCCC+ CCC CCCCC C DDD DD D Aaa Aa1 Aa2 Aa3 A1 A2 A3 Baa1 Baa2 Baa3 Ba1 Ba2 Ba3 B1 B2 B3 Caa1 Caa2 Caa3 23 22 21 20 19 18 17 16 15 14 13 12 11 10 9 8 7 6 5 4 3 2 1 0 D Ca C Source: Author’s transformation, based on Rating agencies methodologies and Gaillard (2009).. Data for macroeconomic variables comes from the World Development Indicators and data on national debt from the Global Development Finance (World Bank). Table 3 provides a résumé of the main macroeconomic variables used across the different rating models. 9 Table 3. Descriptive Statistics for Variables Variables GDP (current US$) GDP (constant 2000 US$) GDP growth (annual %) GDP per capita (constant 2000 US$) GDP per capita, PPP (current international $) Inflation, consumer prices (annual %) Inflation, GDP deflator (annual %) Consumer price index (2000 = 100) Real effective exchange rate index (2000 = 100) Official exchange rate (LCU per US$, period average) Changes in net reserves (BoP, current US$) Current account balance (% of GDP) Current account balance (BoP, current US$) Exports of goods and services (BoP, current US$) Exports of goods, services and income (BoP, current US$) Foreign direct investment, net (BoP, current US$) Foreign direct investment, net inflows (% of GDP) Imports of goods and services (BoP, current US$) Imports of goods, services and income (BoP, current US$) Total reserves (% of external debt) Total reserves (includes gold, current US$) Total reserves in months of imports Total reserves minus gold (current US$) Workers' remittances, receipts (BoP, current US$) Net capital account (BoP, current US$) Bank capital to assets ratio (%) Bank nonperfoming loans to total gross loans (%) Risk premium on lending (%) S&P/EMDB indexes (annual % change) Export quantum/quantity index (2000 = 100) Import value index (2000 = 100) Gross National Product Ratings Fitchs Ratings Moodys Ratings Standard and Poors Fiscal Budget Workers Remittances Workers Remittances / GDP Volatility of GDP Growth Volatility of External flows (incl. remittances) Volatility of External flows (excl. remittances) Solvency Ratio (Debt / Exports) Solvency Ratio (Debt / Exports) excl. remittances Obs Mean Std. Dev. 1661 1655 1649 1655 1636 1475 1642 1497 772 1590 1441 1437 1441 1441 1441 1419 1463 1441 1441 1408 1545 1412 1545 1076 1073 391 422 562 425 1179 1284 1483 485 603 614 858 1076 1072 1378 1150 1150 1261 1340 7.13E+10 6.62E+10 3.777404 1904.606 4154.623 52.26535 86.95552 87.18809 4489.065 503.812 -1.85E+09 -3.532393 2.39E+08 1.95E+10 2.05E+10 1.50E+09 3.352745 1.89E+10 2.14E+10 46.78096 1.13E+10 3.935548 1.07E+10 1.03E+09 8.98E+07 10.40614 10.42891 8.088065 18.53672 94.06167 95.04032 6.84E+10 9.735547 9.972587 9.74107 -2.447704 1.03E+09 0.0369062 3.486176 0.0049457 0.0050462 240.9641 247.5379 2.01E+11 1.73E+11 6.394133 1673.767 3159.793 337.7307 642.4122 54.11652 121757.5 1759.692 1.26E+10 -7.884468 1.05E+10 5.54E+10 5.80E+10 5.16E+09 10.4849 4.85E+10 5.26E+10 125.9236 6.08E+10 3.113531 6.00E+10 2.37E+09 9.64E+08 3.903248 8.365045 14.34093 47.35008 51.80754 48.91408 1.81E+11 3.015993 3.324188 2.983933 4.318253 2.37E+09 0.0476944 4.01902 0.0309158 0.030922 321.0831 319.7211 Sources: Global Development Finance, World Development Indicators, International Financial Statistics, 2009. 3.1.2. Testing Previous Models for Sovereign Ratings: The Effect of Remittances We start by testing the four representative models proposed in the literature, for the period 1993-2006. We expect to identify the most relevant determinants for ratings in our sample. Regarding previous studies, the sample includes a larger number of countries and a 14-year time span. We run regressions on OLS and fixed-effect panel data, using the sovereign rating of the three rating agencies as the dependent variable4. Moreover, we are interested in analysing the impact that remittances could have on the behaviour of rating agencies. As presented, remittances flows can be shock absorbers for the economy, and they play a role in reducing the country’s vulnerability by providing resources that can have a positive impact on the balance of payments and on the appreciation of the exchange rate5. More generally, remittances can improve a country’s creditworthiness and thereby facilitate its access to international capital markets. One significant and key solvency ratio used in the literature to explain the behaviour of sovereign ratings is the debt-to-exports ratio. We start by including this variable in the standard rating models introducing remittances in the ratio’s denominator to capture the entire effect of the current account incomes. At this point we only introduce remittances through the solvency ratio as our second core variable (i.e., volatility of external flows) is not studied in the literature. An increase in the remittances (as well as for the case of the exports of goods and services) can have a positive effect on the current account and can induce appreciations on the real exchange rate. By the same, these revenues in the balance of payments can serve as a cushion against external shocks and then reduce the risk of default of the external debt. In fact, since we are interested in the country’s capacity to pay the entire total external debt (private and public), it is reasonable to include remittances in this ratio, to capture total incomes received by nationals in the balance of payments. To quantify the impact of remittances on sovereign ratings, we test the standard models for ratings by excluding/including the flow of remittances in the external debt to exports ratio. Data on exports comes from the Global Development Finance (GDF) and data on workers´ remittances from the International Financial Statistics (IFS)6. Figure 4 exhibits the evolution of our solvency ratio (debt over exports) for Latin American and Caribbean countries, where the relative impact of remittances in indebtedness indicators remains heterogeneous. In general, the effect of remittances is higher in Central American and Caribbean 4 OLS estimations are not reported and we can provide them under request. However as noted beforehand, remittances may increase moral hazard problems for government public finances and reduce the competitiveness for other sectors of the economy (i.e. Dutch disease effect). 6 Data on exports from the Global Development Finance also include total workers’ remittances registered in the Balance of Payments. The GDF defines Exports of Goods, Services and Income (XGS) as the total value of goods and services exported, and receipts of compensation of employees, and investment income. In order to calculate our solvency ratio we first exclude workers’ remittances and compensation of employees from the XGS variable (solvency ratio without remittances) and then we include workers’ remittances (from the IFS) in the denominator of the solvency ratio (solvency ratio with remittances). Workers’ remittances, a transfer and not an income entry in the balance of payments, are treated as compensation of employees in Global Development Finance because they are often uneasy to distinguish from compensation of non-resident workers and migrants. We therefore have usually workers’ remittances and compensation of employees contained in the Export series. Workers’ remittances and compensation of employees comprise current transfers by migrant workers, wages and salaries earned by nonresident workers. In addition, migrants’ transfers, a part of capital transfers, are treated as workers’ remittances in Global Development Finance. We therefore restrict our analysis to the series of “workers remittances”, and exclude compensation of employees and migrants’ tranfsfers (as estimated by GDF database). See GDP Volume 1 for more details. 5 11 countries (e.g., El Salvador, Jamaica, Guatemala, Dominican Republic) than in other countries of the region (e.g., Argentina, Brazil, Chile, Mexico, Peru, Venezuela). Figure 4. Ratio of External debt over Exports (tdoverx_wr) and over Exports and Remittances (tdoverx) Latin American and Caribbean countries Argentina Belize 800 Bolivia 250 450 700 400 200 600 500 350 300 150 250 400 200 100 300 150 200 50 100 100 50 tdoverx tdoverx_wr tdoverx tdoverx tdoverx_wr Bolivia 250 600 200 500 400 150 300 100 200 50 100 tdoverx tdoverx_wr tdoverx Costa Rica tdoverx 2006 2005 2004 2003 2002 2001 2000 2006 2005 2004 2003 2002 2001 2000 2006 2005 2004 2003 2002 2001 2000 tdoverx tdoverx_wr El Salvador Ecuador tdoverx_wr Grenada 250 350 1999 1993 2006 2005 2004 2003 2002 2001 2000 1999 1998 1997 1996 1995 1994 1993 tdoverx 1998 0 tdoverx_wr 400 1999 20 0 2006 2005 2004 2003 2002 2001 2000 1999 1998 1997 1996 0 40 1997 50 1996 20 60 1995 100 1999 80 200 1995 1998 100 250 150 1997 Dominican Republic 100 40 tdoverx_wr 120 300 60 1996 tdoverx 120 80 1995 tdoverx_wr Dominica 140 1994 1993 2006 2005 2004 2003 2002 1993 tdoverx_wr 2001 0 2006 2005 2004 2003 2002 50 1994 tdoverx 2001 2000 1999 1998 1997 1996 1995 0 2000 50 100 1999 100 1998 150 150 1997 200 1996 250 1994 1997 Colombia 200 1995 300 tdoverx_wr 250 1994 350 1993 1996 tdoverx 200 180 160 140 120 100 80 60 40 20 0 400 1994 1995 tdoverx_wr Chile Brazil 450 1993 1994 1993 2006 2005 2004 2003 2002 2001 2000 1999 1998 1997 1996 1995 1994 1993 2006 2005 2004 2003 2002 0 1998 tdoverx 2001 2000 1999 1998 1997 1996 1995 1994 1993 0 350 300 200 300 250 150 250 200 200 150 100 150 100 100 50 50 50 0 tdoverx tdoverx_wr tdoverx tdoverx_wr Source: Global Development Finance and International Financial Statistics, 2009. tdoverx tdoverx_wr 2006 2005 2004 2003 2002 2001 2000 1999 1998 1997 1996 1995 1994 1993 2006 2005 2004 2003 2002 2001 2000 1999 1998 1997 1996 1995 1994 0 1993 2006 2005 2004 2003 2002 2001 2000 1999 1998 1997 1996 1995 1994 1993 0 2006 2005 2004 2003 2002 2001 2000 1999 1998 1997 1996 1995 1994 1993 1992 tdoverx_wr Belize Argentina 500 450 400 350 300 250 200 150 100 50 0 1991 1990 1987 1989 0 2006 2005 2004 2003 2002 2001 2000 1999 1998 1997 1996 1995 1994 1993 1992 1991 1990 1989 1988 1987 2006 2005 2004 2003 2002 2001 2000 1999 1998 1997 1996 1995 1994 1993 1992 1991 1990 1989 1988 1987 0 1988 0 Figure 4. Ratio of External debt over Exports (tdoverx_wr) and over Exports and Remittances (tdoverx) Latin American and Caribbean countries (cont.) Guatemala Haiti Guyana 3500 350 3000 120 300 100 250 80 200 60 150 40 100 20 50 500 0 0 0 2500 2000 1500 140 tdoverx_wr tdoverx_wr tdoverx 2006 2005 2004 2003 2002 2001 120 3000 100 2500 2000 100 1500 2006 2005 2004 Panama 3500 150 tdoverx_wr tdoverx Nicaragua 200 2003 1993 2006 2005 2004 2003 2002 2001 2000 1999 1998 1997 1996 1995 1994 0 Mexico 250 2000 50 1993 2006 2005 2004 2003 2002 2000 1998 1997 1996 1994 2001 0 1999 0 1995 20 1999 100 40 50 2002 100 1998 150 60 2001 80 2000 150 200 1999 100 1998 200 1997 120 1996 250 1997 250 160 300 1996 Latin America and Caribbean 300 1995 180 1994 200 tdoverx_wr tdoverx Jamaica Honduras 350 1993 1995 tdoverx_wr tdoverx 400 tdoverx 1994 1993 2006 2005 2004 2003 2002 2001 2000 1999 1997 1996 1995 1994 tdoverx_wr tdoverx 1998 1000 1993 2005 2004 2002 2001 2000 1999 1998 1997 1996 1995 1994 1993 2006 400 140 2003 160 80 60 2004 2005 2006 2004 2005 2002 2001 2000 1999 1998 1997 1996 1995 1994 2006 2005 2004 2003 2002 2003 100 2003 120 tdoverx_wr tdoverx Peru 140 2002 Paraguay 2001 2000 1999 tdoverx_wr tdoverx tdoverx_wr tdoverx 1998 1997 1996 1995 1994 0 1993 20 0 2006 2005 2004 2003 2002 2001 2000 1999 1998 1997 1996 1995 1994 1993 0 500 1993 40 1000 50 Uruguay 600 400 500 350 300 400 250 80 60 40 300 200 200 150 20 100 0 0 100 tdoverx_wr tdoverx tdoverx_wr tdoverx Venezuela 250 200 150 100 50 tdoverx 2006 2005 2004 2003 2002 2001 2000 1999 1998 1997 1996 1995 1994 1993 0 tdoverx_wr Source: Global Development Finance and International Financial Statistics, 2009. 13 tdoverx tdoverx_wr 2006 2001 2000 1999 1998 1997 1996 1995 1994 0 1993 2006 2005 2004 2003 2002 2001 2000 1999 1998 1997 1996 1995 1994 1993 2006 2005 2004 2003 2002 2001 2000 1999 1998 1997 1996 1995 1994 1993 50 Following the literature review, we opt for testing our hypothesis on a group of models on sovereign ratings. Table 4 summarizes results of four representative models (Cantor and Packer (1996), Rowland and Torres (2004), Sutton (2005) and Mora (2006)), for the three main agencies over the period 19932006, using both ratios, total debt over exports of goods and services, and workers’ remittances ( TDX ) and total debt over exports of goods and services ( TDX _ wr ). Table 4a. Determinants of Sovereign Ratings (1993-2006) Cantor and Packer (1996), Rowland and Torres (2004) S&P gdp_per_cap_ppp_curr gdp_growth_annual inflat_annual fisc_budget_our current_account_bal_%gdp tdoverx default_20 Cantor and Packer Moodys (iii) (iv) (i) (ii) With Remittances Without Remittances With Remittances 0.0011*** [12.59] -0.0729*** [3.61] -0.0012* [1.92] 0.0184 [0.66] -0.1543*** [8.75] -0.0129*** [7.58] -0.8783** [2.27] 0.0011*** [12.44] -0.0720*** [3.55] -0.0012* [1.88] 0.0193 [0.69] -0.1535*** [8.64] -0.0116*** [7.18] -0.7872** [2.03] Fitch S&P Rowland and Torres Moodys (ix) (x) (v) (vi) (vii) (viii) Without Remittances With Remittances Without Remittances With Remittances Without Remittances With Remittances 0.0012*** [15.38] -0.0516*** [3.20] -0.0003 [0.91] -0.0427* [1.81] -0.1109*** [7.77] -0.0088*** [6.25] -0.3678 [1.14] 0.0012*** [15.23] -0.0507*** [3.13] -0.0004 [0.93] -0.0415* [1.75] -0.1105*** [7.69] -0.0076*** [5.86] -0.3079 [0.95] 0.0011*** [10.58] -0.0467** [2.27] -0.001 [1.25] -0.1157*** [3.22] -0.0774*** [4.53] -0.0071*** [3.66] 0.0096 [0.02] 0.0011*** [10.58] -0.0468** [2.28] -0.001 [1.23] -0.1145*** [3.19] -0.0783*** [4.58] -0.0065*** [3.56] 0.0755 [0.15] -0.0086 [0.40] -0.0012 [1.60] -0.0086 [0.40] -0.0012 [1.60] 0.0089*** [3.20] -2.3087*** [5.32] -0.0581*** [7.68] 0.0491* [1.86] 0.0375** [2.38] 0.0025 [0.31] 417 46 0.57 418 46 0.56 334 47 0.45 335 47 0.45 497 55 0.27 tdovergnp tdtdsovergnp res_gnp openness Fitch (xi) (xii) Without Remittances With Remittances Without Remittances -0.0023 [0.12] -0.0010** [1.98] -0.003 [0.15] -0.0010* [1.96] -0.0466** [2.30] -0.0003 [0.41] -0.0478** [2.36] -0.0003 [0.44] 0.0093*** [3.62] -2.3789*** [5.50] -0.0602*** [7.99] 0.0479* [1.82] 0.0385** [2.45] 0.0035 [0.45] 0.0070*** [2.66] -1.3330*** [3.19] -0.0385*** [5.70] 0.0374 [1.56] 0.0538*** [3.77] -0.0047 [0.66] 0.0066*** [2.75] -1.3697*** [3.29] -0.0391*** [5.75] 0.0391 [1.64] 0.0547*** [3.84] -0.0045 [0.64] 0.0180*** [6.11] -0.3488 [0.67] -0.0734*** [9.70] 0.0384 [1.45] 0.0596*** [3.05] 0.0178** [2.19] 0.0164*** [6.08] -0.5101 [0.97] -0.0727*** [9.68] 0.0435 [1.64] 0.0614*** [3.14] 0.0170** [2.10] 497 55 0.27 477 49 0.22 477 49 0.23 363 50 0.33 363 50 0.33 log_res log_bank_res_to_bank_assets log_tdtdsoverx log_tdovergnp Observations 427 428 Number of country_id 49 49 R-squared 0.53 0.52 Absolute value of t statistics in brackets * significant at 10%; ** significant at 5%; *** significant at 1% Table 4b. Determinants of Sovereign Ratings (1993-2006) Sutton (2005) Sutton Moodys S&P Fitch (xiii) (xiv) (xv) (xvi) (xvii) (xviii) With Remittances Without Remittances With Remittances Without Remittances With Remittances Without Remittances -2.2285*** [5.64] -2.2598*** [5.73] -1.2746*** [3.43] -1.3135*** [3.53] -0.3887 [0.81] -0.4226 [0.88] 1.9091*** [9.06] -0.9255*** [4.65] 0.7199*** [3.45] -1.7545*** [6.34] 361 49 0.42 1.9046*** [9.03] -0.9210*** [4.61] 0.7167*** [3.42] -1.7517*** [6.33] 361 49 0.42 gdp_per_cap_ppp_curr gdp_growth_annual inflat_annual fisc_budget_our current_account_bal_%gdp tdoverx default_20 tdovergnp tdtdsovergnp res_gnp openness log_res 1.8452*** 1.8383*** 1.6582*** 1.6515*** [9.72] [9.69] [10.46] [10.39] log_bank_res_to_bank_assets -0.8874*** -0.8789*** -0.6375*** -0.6345*** [5.33] [5.27] [4.48] [4.44] log_tdtdsoverx 0.8156*** 0.8232*** 0.9460*** 0.9063*** [4.07] [4.13] [5.23] [4.98] log_tdovergnp -1.7675*** -1.7797*** -1.3803*** -1.3590*** [6.43] [6.46] [5.74] [5.63] Observations 488 488 479 479 Number of country_id 55 55 50 50 R-squared 0.4 0.4 0.38 0.37 Absolute value of t statistics in brackets Absolute value of t statistics in brackets * significant at 10%; ** significant *atsignificant 5%; *** significant at 10%; **atsignificant 1% at 5%; *** significant at 1% 15 Table 4c. Determinants of Sovereign Ratings (1993-2006) Mora (2006) Mora (level variable) Moodys Fitch (i) (ii) (iii) (iv) (v) (vi) With Without With Without With Without Remittances Remittances Remittances Remittances Remittances Remittances gdp_per_cap_ppp_curr 0.0012*** 0.0012*** 0.0014*** 0.0015*** 0.0010*** 0.0010*** [10.34] [10.37] [14.06] [14.18] [6.52] [6.50] gdp_growth_annual -0.0877*** -0.0871*** -0.0741*** -0.0746*** -0.0574** -0.0587** [4.57] [4.55] [4.40] [4.46] [2.54] [2.60] inflat_annual -0.0012 -0.001 -0.0005 -0.0005 -0.0034 -0.0039 [0.18] [0.15] [0.62] [0.62] [0.32] [0.37] fisc_budget_our 0.0353 0.0358 -0.0645** -0.0653** -0.0141 -0.0158 [1.22] [1.24] [2.54] [2.58] [0.35] [0.39] current_account_bal_%gdp -0.0931*** -0.0929*** -0.0788*** -0.0796*** -0.0708*** -0.0713*** [4.96] [4.95] [4.67] [4.74] [3.02] [3.03] tdoverx -0.0041* -0.0039* -0.0041** -0.0045** 0.0036 0.003 [1.87] [1.87] [2.07] [2.40] [1.42] [1.22] default_20 -0.4623 -0.4338 -0.135 -0.1095 0.4407 0.4112 [0.93] [0.87] [0.31] [0.26] [0.73] [0.68] spread -0.0010*** -0.0010*** -0.0003*** -0.0003*** -0.0011*** -0.0011*** [8.38] [8.41] [3.30] [3.19] [7.62] [7.54] Observations 225 226 229 230 194 195 Number of country_id 27 28 27 28 25 26 R-squared 0.76 0.76 0.72 0.72 0.69 0.69 Absolute value of t statistics in brackets * significant at 10%; ** significant at 5%; *** significant at 1% S&P Mora (lagged variable) Moodys Fitch (vii) (viii) (ix) (x) (xi) (xii) With Without With Without With Without Remittances Remittances Remittances Remittances Remittances Remittances 0.0001 0.0002 0.0002*** 0.0011 -0.0022 0.0001 [1.18] [1.22] [2.99] [0.65] [4.68] [2.43] 0.1089*** 0.1098*** 0.0001 0.0680*** 0.0001 -0.3482 [5.11] [5.17] [1.97] [3.01] [0.76] [4.72] -0.0082 -0.0079 0.0011 0.0001 0.0008 0.0859*** [1.10] [1.06] [0.07] [2.15] [0.52] [1.89] 0.0199 0.0206 -0.0033* 0.0541 0.0776** 0.0011 [0.62] [0.64] [1.44] [0.15] [0.10] [0.94] 0.0549*** 0.0553*** 0.0206 0.0207 0.0273 0.0275 [2.64] [2.66] [1.44] [1.45] [1.44] [1.45] -0.0041* -0.0037 0.0462** 0.0463** -0.3658 -0.0019 [1.67] [1.60] [4.75] [1.43] [1.06] [0.13] 0.0814 0.1105 0.0677*** 0.0002*** 0.0852*** 0.0784** [0.15] [0.20] [0.60] [4.78] [2.40] [0.52] 0.0005*** 0.0004*** 0.0263 -0.0032** 0.0002* 0.0002* [3.29] [3.26] [2.15] [2.01] [1.94] [0.72] 225 226 229 230 194 195 27 28 27 28 25 26 0.35 0.35 0.33 0.33 0.37 0.37 S&P Results in Table 1 show that for most models the ratio debt over exports (with or without remittances) is negative and significant for the three major rating agencies. Indeed, it is a key and relevant variable explaining sovereign ratings. For instance, taking Cantor and Packer (1996) model, columns 1 to 6 in Table 4a show that the solvency ratio (i.e. Debt over Exports) is statistically significant at 1 per cent and negatively correlated with sovereign ratings. In addition to this ratio, other variables are crucial to explain ratings assigned by agencies: GDP per capita, GDP growth, inflation rate, current account over GDP, historical default and international reserves over GNP. Finally, when comparing the impact of including and excluding remittances on the ratio debt over exports, the value of the coefficient is almost the same for both cases (in absolute terms), for all rating models studied. Results suggest that rating agencies are not sensitive to workers’ remittances when they rate sovereigns. Indeed, an inclusion of remittances supposes a reduction of the solvency ratio and consequently a higher coefficient (in absolute value) can compensate for the “remittances effect” in the sovereign rating. This finding is explained empirically in detail for our general model. 3.1.3. Proposing a General Model and Testing Effect from Remittances Traditional models on the determinants of ratings include a solvency indicator, such as the ratio debt to exports. By introducing remittances (as suggested by Ratha, 2005), we have tested if remittances flows play a role in reducing external vulnerabilities. In addition, we introduce a consistent explanatory variable for sovereign ratings in which remittances can play a crucial role: the volatility of external flows. As specified above, when compared to other external flows (i.e. exports, portfolio flows, FDI flows, ODA), remittances display a much lower volatility and lower correlation to these flows; they can act as a cushion vis-à-vis capital flights. We assess the volatility of external flows as a second channel through which remittances flows are likely to affect sovereign ratings. Our hypothesis is that remittances can reduce the total volatility of inward external flows, which is itself a powerful explanatory variable for sovereign ratings. We use the variance as a measure of volatility of external flows. We decompose the variance of inward external flows as follows: N N i 1 i j Var external _ flows ,t wi2,t Var ( X i ,t ) 2 wi ,t w j ,t Cov( X i ,t , X j ,t ) where Var (external _ flows) ,t corresponds to the variance of inward external flows of country at time t, wi ,t is the weight of the external flow i with respect to the total external flows in country , Var ( X i ,t ) is the variance of the external flow i as a share of GDP between t-5 and t, Cov ( X i ,t , X j ,t ) is the covariance between the external flows over GDP i and j and from t-5 to t. Figure 5 presents the volatility of external flows by including and excluding remittances. There is a considerable reduction of the external flows volatility for some Central American countries (e.g., Honduras, Guatemala, Nicaragua) and high-remittance receptors (e.g. Ecuador, Colombia, El Salvador). 17 Figure 5a. Volatility of Inward External Flows with and without Workers’ Remittances Bolivia Argentina Brazil 0.0035 0.0045 Volat External Flows 0.004 Volat External Flows 0.00035 0.0025 0.003 0.002 0.002 0.0015 0.0015 Volat External Flows (without remitt.) 0.0003 Volat External Flows (without remitt.) 0.0025 Volat External Flows 0.0004 0.003 Volat External Flows (without remitt.) 0.0035 0.00045 0.00025 0.0002 0.00015 0.001 0.001 0.0001 0.0005 0.00005 0 Ecuador Dominican Repubic 0.0025 0.002 0.0014 Volat External Flows (without remitt.) 0.0012 0.0015 0.0008 0.001 0.0006 Peru 0.0012 Volat External Flows (without remitt.) 2005 2007 2004 2005 2006 2003 2004 2002 2001 2000 1999 2003 2007 2006 2002 2001 2007 2006 2005 2004 2003 2002 2001 2000 1999 2007 2006 2005 2004 2003 2002 2001 2000 1999 1997 1996 1995 1994 2007 2006 2005 2004 2003 2002 2001 2000 1997 1996 1995 1994 1993 2007 2006 2005 2004 2003 2002 2001 Rating Rating Venezuela 0.014 Volat External Flows 0.012 Volat External Flows (without remitt.) 0.01 0.008 0.0015 0.0006 1993 2007 2006 2005 2004 2003 2002 2001 2000 1999 2000 1999 1998 Volat External Flows (without remitt.) 0.002 0.0008 Volat External Flows (without remitt.) Rating 0.0025 0.001 Volat External Flows 0.016 Volat External Flows Volat External Flows 1998 1997 1996 1995 1994 1993 2007 2006 2005 2004 2003 2002 2001 2000 1999 1998 1997 1996 1995 Uruguay 0.003 0.0014 Paraguay Rating 0.0016 1993 2007 2006 2005 2004 2003 2002 2001 2000 1999 1998 1997 1996 1995 1994 0 1997 0.01 0.005 Volat External Flows (without remitt.) 1996 Rating 0.015 Volat External Flows (without remitt.) 0.009 0.008 0.007 0.006 0.005 0.004 0.003 0.002 0.001 0 Volat External Flows 1995 Volat External Flows (without remitt.) Panama 1994 Volat External Flows 1994 2007 2002 1993 0 2006 0.001 0 2005 0.002 0.001 2004 0.003 0.002 2003 0.004 0.003 2001 0.004 2000 0.005 1999 0.006 0.005 1998 0.006 1997 0.007 0.1 0.09 0.08 0.07 0.06 0.05 0.04 0.03 0.02 0.01 0 Volat External Flows 0.007 Volat External Flows (without remitt.) 1996 1995 1994 1993 Mexico Honduras Volat External Flows Volat External Flows 0.02 1993 1998 1997 1996 1995 1994 1993 2007 2006 2005 2004 2003 2002 2001 2000 1999 1998 1997 1996 1995 1994 1993 0 Guatemala Nicaragua 2000 0.0002 0 0.025 1999 Volat External Flows (without remitt.) 0.0004 0.0005 0.0009 0.0008 0.0007 0.0006 0.0005 0.0004 0.0003 0.0002 0.0001 0 1998 1997 1996 Volat External Flows 0.001 1999 Volat External Flows (without remitt.) El Salvador 0.0016 Volat External Flows 1998 Volat External Flows 1998 1993 2007 2006 2005 2004 2003 2002 2001 2000 1999 1998 1997 1996 1994 1993 2007 2006 2005 2004 2003 2002 2001 2000 1999 1998 1997 1996 1995 1994 1993 1995 0 0 1998 0.0002 1997 0.00005 1996 0.0001 0.0004 1995 0.00015 0.0006 1994 0.001 0.01 0.009 0.008 0.007 0.006 0.005 0.004 0.003 0.002 0.001 0 1995 1993 2007 2006 2005 2004 2003 2002 2001 2000 0.0002 0.0008 Volat External Flows 0.0045 0.004 0.0035 0.003 0.0025 0.002 0.0015 0.001 0.0005 0 Volat External Flows (without remitt.) Volat External Flows (without remitt.) 0.0012 Costa Rica Volat External Flows 0.00025 Volat External Flows 0.0014 1999 Colombia Chile 0.0016 1998 1997 1996 1995 1994 1993 0 2007 2006 2005 2004 2003 2002 2001 2000 1999 1998 1997 1996 1995 1993 1994 0 1994 0.0005 0.006 0.001 Source: Global Development Finance and International Financial Statistics, 2009. 2007 2006 2005 2004 2003 2002 2001 2000 1999 1998 1997 1996 1995 1994 2007 2006 2005 2004 2003 2002 2001 2000 1999 1998 1997 1996 0 1995 0.002 0 1994 2007 2006 2005 2004 2003 2002 2001 2000 1999 1998 1997 1996 1995 1994 1993 0 0.0005 1993 0.0002 1993 0.004 0.0004 Figure 5b. Average Volatility of External Flows with and without Workers’ Remittances 0.007 Impact of Remittances on Volatility of External Flows 0.006 Average 1992-2006 0.005 Average Volat External Flows 0.004 Volat External Flows (without remitt.) 0.003 0.002 0.001 Colombia Brazil Guatemala El Salvador Peru Bolivia Chile Ecuador Uruguay Mexico Argentina Costa Rica Honduras Dominican Republic Venezuela, RB Paraguay Nicaragua 0 Source: Authors’ calculation, based on Global Development Finance and International Financial Statistics, 2009. Considering the results from the four standard models and our new volatility indicator presented above, we propose the following model for our analysis (we name it General Model): Rating i ,t 0 1GDP _ pc 2GDP _ growthi ,t 3 Inflat i ,t 4 Fisc _ budg i ,t 5CAi ,t 6TDX i ,t 7 Default i ,t 8 Re servesi ,t 9Volat _ indicatori ,t 7 EMBI i ,t t i i ,t where Rating i ,t corresponds to the transformed rating of country i at time t, GDP _ pc is the per capita GDP in current international dollars, GDP _ growth is the product’s growth, Inflat corresponds to annual inflation, Fisc _ budg is the annual balance budget as a share of GDP, CA is the current account position (as a share of GDP), TDX is the ratio of total debt to exports, Default is a dummy variable for countries taking value 1 for countries having experienced a default during the previous 20 years, Re serves is the ratio of reserves to GNP, Volat _ indicator is the volatility of external of flows, EMBI is a dummy variable for those countries covered by the Bond Index calculated by JPMorgan, t is a year fixed effect, i is the country-individual effect and i,t is an error term. Within this setup, 6 measures the elasticity of sovereign ratings with respect to the ratio debt-to-exports, after controlling for all the other factors (GDP, inflation, volatility, etc.). The term t is capturing differences in sovereign rating across time not explained by the other determinants. 19 In this model, we are mainly interested in the variables that can be affected by the flow of remittances, particularly the volatility indicator (i.e., the volatility of inward external flows) and the solvency ratio (i.e., debt-to-exports). 7 Table 5 shows the results for our general model and the three rating agencies. We report results using the two variables of interest, the volatility indicator and the solvency ratio, including and excluding remittances8. Table 5. General Model – All Sample S&P GDP per capital (PPP) GDP growth (annual) Annual inflation Fiscal Budget Current Account (% GDP) Solvency Ratio (debt/exports) Default dummy (20 years) Reserves Ratio Volatility External Flows EMBI dummy Observations Number of country_id R-squared Standard errors in brackets *** p<0.01, ** p<0.05, * p<0.1 (i) With remittances 0.0010*** [0.0001] -0.0249 [0.0219] -0.0013** [0.0006] -0.0204 [0.0332] -0.1330*** [0.0208] -0.0121*** [0.0018] -1.2455*** [0.4294] 0.0091 [0.0167] -155.7120*** [37.6048] 0.5923** [0.2979] 360 43 0.528 Moodys ME 4.1235 -0.0944 -0.0632 0.0498 0.4714 -2.6658 -0.9099 0.1401 -0.7552 0.1213 (ii) Without Remittances 0.0010*** [0.0001] -0.03 [0.0202] -0.0012** [0.0006] -0.0149 [0.0290] -0.1375*** [0.0178] -0.0111*** [0.0016] -1.1585*** [0.4079] 0.0058 [0.0144] -131.2598*** [28.3887] 0.6037** [0.2595] 398 47 0.535 ME 4.1235 -0.1138 -0.0584 0.0364 0.4873 -2.4455 -0.8464 0.0893 -0.6366 0.1237 (iii) With remittances 0.0012*** [0.0001] -0.0540*** [0.0177] -0.0003 [0.0004] -0.0743*** [0.0282] -0.1031*** [0.0176] -0.0084*** [0.0015] -0.4639 [0.3641] 0.0335** [0.0139] -7.4536* [3.8138] 0.0683 [0.2509] 353 39 0.571 ME 4.9482 -0.2048 -0.0146 0.1815 0.3654 -1.8506 -0.3389 0.5156 -0.0361 0.0140 Fitch (iv) Without Remittances 0.0011*** [0.0001] -0.0536*** [0.0170] -0.0003 [0.0004] -0.0453* [0.0257] -0.1140*** [0.0154] -0.0072*** [0.0014] -0.3248 [0.3577] 0.0249** [0.0127] -8.7014** [3.7647] 0.0673 [0.2442] 390 44 0.556 ME 4.5359 -0.2032 -0.0146 0.1107 0.4040 -1.5863 -0.2373 0.3832 -0.0422 0.0138 (v) With remittances 0.0010*** [0.0001] -0.0126 [0.0229] -0.0011 [0.0008] -0.1141** [0.0495] -0.0840*** [0.0187] -0.0056** [0.0022] -0.2517 [0.6137] -0.0176 [0.0270] -173.4265*** [42.6915] 0.4242 [0.4211] 273 41 0.455 ME 4.1235 -0.0478 -0.0535 0.2788 0.2977 -1.2338 -0.1839 -0.2709 -0.8411 0.0869 (vi) Without Remittances 0.0010*** [0.0001] -0.0119 [0.0215] -0.001 [0.0008] -0.1288*** [0.0424] -0.0842*** [0.0178] -0.0053*** [0.0020] -0.3075 [0.5871] -0.0214 [0.0234] -115.0774*** [32.8760] 0.3523 [0.3655] 305 45 0.439 Source: Authors’ calculation. First, we analyze results for regressions including remittances in the volatility of capital flows as well as in the debt ratio. We focus on regressions (i), (iii) and (v) for Table 5. Not surprisingly, GDP per capita is positive and significant at the 1 per cent level. GDP growth, on the contrary, is not significant for our sample and it is negatively correlated with ratings (the exception being for Moody’s). An increase of 7 As for the case of the models presented above, we build an artificial ratio by subtracting the total amount of Workers’ Remittances from the variable Total Exports, and we name it TDX _ wr , this is, the ratio debt-to-exports excluding workers’ remittances. Again, when comparing the coefficient for the variable Total Debt/Exports with and without remittances, these are very similar. A coefficient test shows that they are not significantly different from the previous regression. 7 By the same, in order to analyse the impact of remittances through the volatility of capital flows, we subtract Workers’ Remittances to the calculation of the volatility of external flows, and we name it Volat _ indicator _ wr , this is, the volatility of external flows by excluding workers’ remittances. 8 A complementary approach consisted of defining a variable taking the difference between the two ratios and volatilities, this is: Debt Debt 1 Exports Exports remit tan ces and 2 Volatwith _ remit Volatwith out_ remit , and test for the significance of these variables in the general model. For the first difference, when including them in the model together with the ratio debt over exports it was significant at 1%, but it becomes non-significant when excluding the ratio. The second difference is not significant in the model at the 5% level (except for Fitch). ME 4.1235 -0.0451 -0.0486 0.3147 0.2984 -1.1677 -0.2247 -0.3294 -0.5581 0.0722 inflation supposes a decrease in the rating, and this result is significant for Standard and Poor’s. The balance budget is negative and significant for the case of Fitch. Unexpectedly, the current account balance is negatively associated to the ratings, indicating that countries running current account deficits may be better rated; there are multiple explanations for this result, one of them being that an important number of commodity exporters, with fragile macroeconomic performance in the past, experienced significant surpluses. Both variables of interest, the debt-to-exports ratio and the volatility of external flows, are consistently negative and significant for all rating agencies. Indeed, additionally to the standard variable used to explain the impact of remittances on ratings (i.e., the solvency ratio), the new variable introduced (i.e., the volatility indicator) is an important one when explaining ratings. The variable default is negatively correlated to the sovereign rating, as expected, and is significant for Standard and Poor’s. The reserves-to-national product ratio is positively related to the rating in the case of S&P and Moody’s, highlighting the increasing role of precautionary reserves for impeding defaults. Now we focus on regressions (ii), (iv) and (vi), that considers the new variables excluding remittances from the debt/exports ratio and the volatility of external flows. The results do not change significantly, and are indeed very similar to the ones using the original indicators. 3.3. Counterfactual analysis for Latin America – General Model To assess the potential effect that the modified solvency ratio as well as the volatility indicator could have on the potential rating for some countries on Latin America, we construct a simple counterfactual scenario, looking at the rating evolution when remittances flows are taken into account. We use the observed ratio debt-to-exports and the counterfactual ratio debt-to-exports that we estimated for the previous regressions, this is, excluding remittances flows (TDX_wr). By the same, we use the observed volatility indicator and the counterfactual volatility indicator that we estimated for the previous regressions, this is, excluding remittances flows (volat_indicator_wr). We estimate our initial model with the counterfactual variable: Rating i ,t 0 1GDP _ pc 2GDP _ growthi ,t 3 Inflat i ,t 4 Fisc _ budg i ,t 5CAi ,t 6TDX _ wri ,t 7 Default i ,t 8 Re servesi ,t 9Volat _ indicatori ,t 7 EMBI i ,t t i i ,t and we obtain the vector ˆ as the fixed-effect estimator. Then, we use the observed ratio debt-toexports (TDX) as well as the volatility of external flows and calculate the change in the observed rating using this ratio and the ˆ coefficients. We obtain the potential improvement in the sovereign ratings for the Latin American countries included in the sample. Figure 6a depicts the observed rating for Standard and Poor’s (“Observed” Y in the figure), the predicted model (“Predicted” Yˆ in the figure) that estimates ratings depending on the observed TDX_wr ratio (debt-to-exports excluding remittances) and on the observed volat_indicator_wr ratio (volatility of external flows excluding remittances) as well as the potential rating in the scenario including workers’ remittances in our variable of interests, debt/exports ~ and volatility of external flows (“Counterfactual” Y in the figure). Figure 6a, 6b, and 6c also provide shadow and predicted ratings for Latin American countries (in the figures, both categories are included under the name “predicted”). The naming “shadow ratings” concerns unrated countries for a given year and a specific agency, while predicted ratings refer to 21 countries that are already assigned a rating (see Table 6 for the two series of periods covering shadow and predicted ratings respectively for each country and agency). Table 6. Shadow and Predicted Ratings by CRA and by Country Argentina Bolivia Brazil Chile Colombia Costa Rica Dominican Rep. Ecuador El Salvador Guatemala Honduras Mexico Nicaragua Panama Paraguay Peru Uruguay Venezuela S&P NA 1993:1997 1993 NA NA 1995:1996 1995:1996 1993:1999 1993:1995 1993:2000 1993:2006 NA 2002:2006 1994:1996 NA 1993:1996 NA NA "Shadow ratings" Moody's NA 1993:1997 NA 1993:1998 NA 1995:1996 1995:1998 1993:1996 1993:2001 1993:1996 1993:1998 NA NA 1994:1996 1995:1997 1993:1998 NA NA Fitch 1993:1997 1993:2003 1993 1993:1995 1993 1995:1997 1995:2002 1993:2001 1993:1995 1993:2005 1993:2006 1993:1995 2002:2006 1994:1997 1995:2006 1993:1998 NA NA S&P 1993:2006 1998:2006 1994:2005 1993:2006 1993:2006 1997:2006 1997:2006 2000:2006 1996:2006 2001:2006 NA 1993:2006 NA 1997:2006 1995:2006 1997:2006 2002:2006 2003:2006 "Predicted ratings" Moody's Fitch 1993:2006 1997:2006 1998:2006 2004:2006 1993:2005 1994:2005 1999:2006 1996:2006 1993:2006 1994:2006 1997:2006 1998:2006 1999:2006 2003:2006 1997:2006 2002:2006 2002:2006 1996:2006 1997:2006 2006 1999:2006 NA 1993:2006 1995:2006 2002:2006 NA 1997:2006 1998:2006 1998:2006 NA 1999:2006 1999:2006 2002:2006 2002:2006 2003:2006 2003:2006 Sources: Fitch (2008a), Moody’s (2008), S&P (2007). Results are very similar from one agency to the other. The main findings are the following. First, ratings yielded by the “predicted” and the “counterfactual” models are, on average, lower than observed ratings for investment grade issuers (Chile, Mexico) and Peru. By contrast, predicted and counterfactual ratings are higher for defaulting countries (Argentina, Dominican Republic, Ecuador). This may prove that the model tends to attenuate the boom-bust debt cycle. Second, when they are compared for a given country, predicted and counterfactual ratings are very close, except for smaller countries (i.e. El Salvador, Honduras, and Nicaragua): the latter being higher than the former. This gap could suggest that ratings of small economies are more likely to be enhanced by the inclusion of remittances9. Source: Authors’ calculation. 23 Uruguay 8 6 Observed Counterf. 0 14 Predicted 4 Observed 2 Counterf. 2007 0 2007 Predicted 2006 6 2006 8 2005 10 2005 10 Venezuela 2004 12 2004 12 2003 2 2003 4 2002 12 2002 Panama 2001 2007 2006 2005 2004 2003 2002 Counterf. 2001 Observed 2 2000 4 2000 14 1999 2007 2006 2005 2004 2003 2002 2001 2000 1999 Counterf. 1999 Counterf. Observed 2 1998 Predicted 4 1998 14 1998 Counterf. 1997 2007 2006 2005 2004 2003 2002 2001 2000 1999 1998 1997 Rating 12 1997 Predicted 1996 10 1997 Observed 1995 9.5 1996 Honduras 1996 0 1996 Counterf. 1995 Predicted 1995 Observed 1995 Ecuador 1994 1993 0 1994 1993 2007 2006 2005 2004 2003 2002 Predicted 1994 Rating 6 1993 Rating 2007 2006 2005 2004 2003 2002 2001 10 1994 1993 Observed Rating 2007 2006 2005 2004 2003 2002 2001 2000 1999 6 2007 2006 2005 2004 2003 2002 2001 2001 2000 1999 Colombia 2001 2 1998 1997 12 2000 4 2007 2006 2005 2004 2003 2002 2001 2000 1999 1998 1997 1996 1995 1994 1993 2007 2006 2005 2004 2003 2002 2001 2000 1999 1998 1997 1996 1995 2 1999 Predicted Rating 12 1998 14 4 1997 0 Counterf. 1996 Counterf. Predicted 1995 Observed Observed 1994 8 Bolivia 1993 10 Rating Peru 8 2007 0 2006 Counterf. 2005 Predicted 2004 Observed 2003 8 2002 10 14 12 10 8 6 4 2 0 -2 -4 -6 -8 2001 Nicaragua 2000 0 2000 Counterf. 2000 Predicted 1999 Observed 10 9 8 7 6 5 4 3 2 1 0 1999 8 Counterf. 1999 Guatemala Observed 2 1998 10 4 1998 0 1998 2 1998 12 1997 Counterf. 1996 14 1997 4 1997 Predicted 1997 Observed 1996 8 1994 1993 Rating 10 9 8 7 6 5 4 3 2 1 0 1996 8 1996 10 1996 Dominican Repubic 1995 10 1995 Counterf. 1995 Predicted 1995 Observed 1994 Chile 1994 1993 0 1994 1993 2007 Counterf. 1994 Rating 2006 2005 2004 2003 2002 2001 2000 Predicted 1993 6 Rating 2007 2006 2005 2004 2003 2002 2001 1999 1998 Observed 1993 6 Rating 2007 2006 2005 2004 2003 2002 2001 2000 1999 1997 6 1995 6 Rating 2007 2006 2005 2004 2003 2002 2001 2000 1999 1998 1996 8 1994 6 Rating 2007 2006 2005 2004 2003 2002 2001 2000 2000 1999 1998 1997 1995 1994 1993 Rating 10 1993 2007 2006 2005 2004 2003 2 2002 4 2001 12 2000 2 1999 4 1999 12 1998 2 1998 4 1998 12 1997 2 1997 4 1997 1996 1995 1994 1993 Rating 18 16 14 12 10 8 6 4 2 0 1997 1996 1995 1994 1993 Rating 12 1996 1995 1994 1993 Rating 2 1996 1995 1994 1993 Rating 4 1996 1995 1994 1993 Rating 12 Argentina Figure 6a. Counterfactual Analysis for Latin America – S&P Brazil 10 8 6 Observed 0 Predicted Counterf. 11.5 Costa Rica 10.5 11 Observed Predicted 9 Counterf. 12 El Salvador 10 8 6 Predicted 0 12 Mexico 10 8 6 Predicted 0 Paraguay 10 8 6 Observed 0 Predicted Counterf. Source: Authors’ calculation. 4 2 0 0 12 12 10 Predicted Counterf. 0 Venezuela 10 8 6 4 2 Observed Predicted 0 Counterf. 2007 2 2006 4 2007 Counterf. 2005 Observed 2006 10 2004 12 2005 12 2003 Panama 2004 2007 2006 2005 2004 Counterf. 2003 Predicted 1 2 2002 4 Counterf. 6 2003 Predicted 2002 2 2001 Predicted 2002 Observed 2001 3 2000 Observed 2001 16 2000 Honduras 1999 2007 2006 2005 2004 2003 2002 2001 2000 14 1999 Ecuador 1998 2007 2006 2005 2004 2003 2002 2001 2000 1999 1998 12 1999 12 1997 Colombia 1998 7 2007 2006 2005 2004 2003 2002 2001 2000 1999 1998 1997 1996 1995 1994 12 2000 Observed 1993 Rating Bolivia 1999 Uruguay Predicted 1998 0 Observed 2 1997 6 4 1998 2 1997 4 1997 10 1996 14 1996 9.5 1997 8 1995 Predicted 1996 Counterf. 1996 Predicted 1995 Observed 1995 0 1995 Counterf. 1994 Observed 1994 10 1996 6 1993 12 1994 0 1995 4 1993 2007 2006 2005 2004 2003 2002 6 1994 6 Rating 8 1993 Rating 2007 2006 2005 2004 2003 2001 2000 4 1993 Predicted Rating 5 Rating 2007 2006 2005 2004 2003 2002 2001 1999 1998 5 2007 2007 2006 2005 2004 2003 2002 2001 2000 1999 7 1994 8 Rating 2006 2005 2004 2003 2002 2002 2001 2000 1999 1997 8 1993 2007 2 2006 Predicted 2005 4 2004 14 2003 Peru 2001 0 6 2002 Counterf. 8 2000 4 2001 14 2000 9 1999 14 1998 Counterf. 1998 Predicted 1 1999 6 1996 2 Counterf. 2000 7 1997 Predicted 1998 Nicaragua 1997 Rating 9 1999 8 1995 Observed 1998 0 1996 3 1997 Counterf. 1994 1993 2007 2006 2005 2004 2003 2002 2001 2000 1999 1998 1997 1996 1995 1994 1993 Moodys Rating Observed 1998 8 1997 10 1996 Guatemala 1997 0 1996 Counterf. 10 9 8 7 6 5 4 3 2 1 0 1996 Predicted 1995 2 1995 Observed 1995 8 1995 10 1994 Dominican Repubic 1994 4 1996 Counterf. 1993 0 1994 12 1995 Observed 1993 Rating 14 1994 6 Rating 2007 2006 2005 2004 2003 2002 8 1993 6 Rating 2007 2006 2005 2004 2003 2002 2001 2000 10 1993 5 Rating 2007 2006 2005 2004 2003 2002 2001 2000 Chile 1994 Rating 2007 2006 2005 2004 2003 2002 2001 2000 1999 16 1993 2007 2006 2005 2004 2003 2002 Predicted 2001 2 2001 Observed 2000 3 2000 9 1999 Counterf. 1999 Predicted 2 1999 1998 4 1998 12 1998 1997 Observed 1998 1997 1996 1995 1994 1993 Rating 18 1999 1 1997 2 1997 4 1996 Rating 6 1998 10 9 8 7 6 5 4 3 2 1 0 1996 1995 1994 1993 2 1996 1995 1994 1993 Rating 4 1997 1995 1994 1993 Rating 12 1996 1995 1994 1993 Rating 10 9 8 7 6 5 4 3 2 1 0 Argentina Figure 6b. Counterfactual Analysis for Latin America – Moody’s Brazil 10 8 6 0 Counterf. Costa Rica 11.5 10.5 11 Observed 9 Predicted Counterf. El Salvador 12 10 8 6 Observed Predicted 0 Counterf. Mexico 10 8 Observed 0 Paraguay 10 8 6 Observed Predicted Counterf. Source: Authors’ calculation. 25 Uruguay Observed 0 6 16 14 12 12 10 10 Venezuela 8 Predicted 4 Predicted 6 4 Observed Predicted Counterf. 2 Counterf. 2 0 Counterf. 2007 16 2007 Observed Counterf. 2006 Panama Observed 2 2005 0 4 2006 Counterf. 2004 Predicted 2005 Observed 2004 14 2003 Counterf. 2003 0 2002 2 2002 Predicted Counterf. 2001 4 2 6 2001 Predicted 2006 2007 2007 2003 2002 2007 2003 2002 2007 2001 2000 1999 1998 1997 1996 1995 1994 1993 2007 2006 2005 2004 2003 2002 2005 0 2006 Predicted 2006 6 2006 8 2004 10 2005 Mexico 2005 12 2005 0 2004 Observed 2004 8 2004 10 2003 El Salvador 2003 0 2002 Observed 2002 8 2001 2001 2000 Rating Costa Rica 2001 2000 Observed 2000 4 2000 Counterf. 1999 2 1999 Predicted 2 1998 4 Counterf. 1998 Predicted 2 1999 12 1997 4 Counterf. 1999 10 1996 Predicted 2000 8 Predicted 1998 8 Observed 2 1998 14 1997 16 1997 Ecuador 1995 4 6 1996 12 1995 0 1994 Observed Observed 1996 10 1993 10 1994 12 10 1995 Honduras 1993 2007 2006 2005 2004 12 1994 6 Rating 8 1993 6 Rating 2007 2006 2005 2004 2003 2001 2000 1999 1998 1997 Rating 12 1999 Rating 2007 2006 2005 2004 2003 2002 16 1997 2007 2006 2005 2004 2003 2002 8 4 1998 Counterf. 2003 2002 2001 14 1996 2 2002 2001 2000 14 1995 Predicted Counterf. 2001 2001 2000 Colombia 1994 1993 4 2000 2000 1999 1998 14 1997 14 Rating Predicted 1999 1999 1998 1996 Counterf. 1996 6 1997 1995 1994 Predicted 1995 0 Observed 1994 Observed Bolivia 1993 8 2007 10 2006 16 2005 Peru 2004 Counterf. 2003 Predicted 2002 Observed 16 14 12 10 8 6 4 2 0 -2 -4 2001 Nicaragua 2000 0 1999 2 1999 12 1998 4 1997 14 1998 Predicted 1996 16 1998 6 1997 10 1997 12 1997 0 1996 Predicted 1996 Counterf. 1995 6 6 1996 Guatemala 1993 Rating 10 9 8 7 6 5 4 3 2 1 0 1996 Observed 1995 8 1995 Dominican Repubic 1995 10 1994 0 1994 8 1993 10 1994 Chile 1993 2007 0 1994 12 16 Rating 2006 2005 2004 2003 2002 2001 2000 Predicted 1993 6 Rating 2007 2006 2005 2004 2003 2002 2001 6 1993 8 Rating 2007 2006 2005 2004 2003 2002 2001 1999 1998 8 1995 Rating 2007 2006 2005 2004 2003 2002 2001 2000 1999 10 1994 6 Rating 2007 2006 2005 2004 2003 2002 2001 2000 1999 1997 Rating 12 1993 2007 2006 2005 2004 2003 2002 12 2001 Counterf. 2000 2 2000 Observed 2000 4 1999 1998 18 1999 1998 1996 Counterf. 1998 1997 1995 1994 1993 2 1998 1997 1996 1995 1994 1993 Rating Observed 1999 2 1997 14 1997 Rating 4 1998 4 1996 10 9 8 7 6 5 4 3 2 1 0 1996 1995 1994 1993 2 1996 1995 1994 1993 Rating 4 1997 1995 1994 1993 Rating 12 1996 1995 1994 1993 Rating 14 Argentina Figure 6c. Counterfactual Analysis for Latin America – Fitch Brazil 10 8 6 0 Counterf. Paraguay 14 12 10 8 Observed 0 Including the debt-to-exports ratio and the volatility of external flows in the estimation does not substantially alter the estimation. We infer that including remittances in the rating agencies’ model does not have an effect on improving the rating for most Latin American countries. To check the robustness of this result, we test the opposite estimation, using the variables TDX and Volat_indicator as follows: Rating i ,t 0 1GDP _ pc 2GDP _ growthi ,t 3 Inflat i ,t 4 Fisc _ budg i ,t 5CAi ,t 6TDX i ,t 7 Default i ,t 8 Re servesi ,t 9Volat _ indicatori ,t 7 EMBI i ,t t i i ,t and doing the counterfactual with the variable excluding remittances (TDX_wr). The results are very similar to those presented in Figures 6a, 6b, and 6c10. 3.3. Model for High-Remittance Receptors The wide range of countries in the sample makes it difficult to identify if there is a clear impact of remittances flows in the regressions. Initially, we would expect that for those countries where remittances have a non-negligible weight (as a share of GDP) in the economy, the change in our two benchmark variables (i.e., solvency ratio and volatility indicator) including and excluding remittances would be significant. For this reason, we decide to calculate a threshold variable (for each country and year) taking the value 1 when the ratio Remittances/GDP was higher than a given percentage and zero otherwise. The objective is to identify those countries and years where remittances are more important. Note that this dummy is non constant over time, and therefore can be included in the fixed-effect panel. Then, we calculate a crossed term with the non-constant dummy and the variables TDX and Volat_indicator, that will detect the interaction effect between countries with a high share of remittances and our variable of interest11. Thus, we test the following model for the whole sample: Rating i ,t 0 1GDP _ pc 2 GDP _ growthi ,t 3 Inflat i ,t 4 Fisc _ budg i ,t 5CAi ,t 6TDX i ,t 7 Default i ,t 8 Re servesi ,t 9Volat _ indicatori ,t 10 EMBI i ,t 11Threshold TDX 12Threshold Volat _ indicatori ,t 13Threshold t i i ,t where Threshold takes the value 1 when the ratio Remittances/GDP was higher than a given percentage and zero otherwise. Threshold TDX and Threshold Volat _ indicator are the interaction effects between countries with a high share of remittances and the ratio debt over exports and the volatility of external flows respectively. 10 11 These results are not reported but they can be provided under request. We tested other configurations to take into account the importance of isolating those ratings most likely to be affected by the remittances flows. We included the ratio remittances-to-GDP as an explanatory variable in equation (1), but this was not significant for the sample. Also, we separated the sample in income groups, following the World Bank classification (lower income/middle income/higher income, etc.) and performed regressions on each group. Finally, we opted for the non-constant dummy variable. Table 7 summarizes the result for the sample, using two different thresholds: 3.5 and 5.0 per cent, respectively. ` Table 7. Regression with Threshold Model Threshold: 3.5% Threshold: 5.0% S&P ME Moodys ME Fitch ME S&P ME Moodys ME Fitch 0.0010*** 4.1235 0.0011*** 4.5359 0.0008*** 3.2988 0.0010*** 4.1235 0.0011*** 4.5359 0.0008*** [9.472] [11.22] [6.283] [8.931] [10.72] [6.057] GDP growth (annual) -0.0385* -0.1460 -0.0390* -0.1479 -0.0244 -0.0925 -0.0375 -0.1422 -0.0385* -0.1460 -0.0137 [-1.652] [-1.965] [-1.012] [-1.569] [-1.913] [-0.553] Annual inflation -0.0013** -0.0632 -0.0002 -0.0097 -0.0013* -0.0632 -0.0013** -0.0632 -0.0002 -0.0097 -0.0014* [-2.177] [-0.595] [-1.698] [-2.139] [-0.542] [-1.677] Fiscal Budget -0.0362 0.0884 -0.1090*** 0.2663 -0.1640*** 0.4007 -0.0217 0.0530 -0.1022*** 0.2497 -0.1693*** [-0.996] [-3.476] [-3.125] [-0.586] [-3.223] [-3.137] Current Account (% GDP) -0.1447*** 0.5128 -0.0879*** 0.3115 -0.0999*** 0.3541 -0.1526*** 0.5408 -0.0890*** 0.3154 -0.0912*** [-6.126] [-4.309] [-4.786] [-6.274] [-4.321] [-4.320] Default dummy (20 years) -1.3282*** -0.9704 -0.5458 -0.3988 -0.2825 -0.2064 -1.0153** -0.7418 -0.4625 -0.3379 -0.3351 [-3.214] [-1.562] [-0.475] [-2.435] [-1.314] [-0.549] Reserves Ratio -0.0043 -0.0662 0.0265* 0.4079 -0.0441 -0.6787 0.0101 0.1554 0.0323** 0.4971 -0.0347 [-0.261] [1.860] [-1.541] [0.603] [2.270] [-1.191] Volatility External Flows -314.7492*** -1.5265 -197.1642*** -0.9562 -248.3374*** -1.2044 -344.3752*** -1.6701 -207.8495*** -1.0080 -225.2765*** [-5.589] [-3.866] [-4.413] [-5.975] [-4.023] [-3.999] EMBI dummy 0.9925*** 0.2033 0.4567 0.0936 0.9060** 0.1856 1.0711*** 0.2194 0.4375 0.0896 0.9737** [3.130] [1.630] [2.012] [3.357] [1.547] [2.092] Threshold dummy x (Debt/exports) -0.0023 -0.5067 0.0011 0.2423 -0.0059 -1.2999 -0.0027 -0.5948 0.0006 0.1322 -0.0125 [-0.480] [0.411] [-1.215] [-0.457] [0.208] [-1.556] Threshold dummy x (Volat. External flows)245.5258** 1.1907 193.0439** 0.9362 123.8197 0.6005 340.2644*** 1.6502 265.8059*** 1.2891 192.1485 [2.108] [2.019] [1.157] [2.776] [2.667] [1.277] Solvency Ratio (debt/exports) -0.0137*** -3.0183 -0.0072*** -1.5863 -0.0032 -0.7050 -0.0141*** -3.1064 -0.0073*** -1.6083 -0.003 [-7.079] [-4.181] [-1.322] [-7.108] [-4.120] [-1.147] Threshold dummy 1.9353*** 0.6517 0.4962 0.1671 2.2812*** 0.7682 1.6900** 0.5691 -0.1591 -0.0536 0.9216 [2.686] [1.041] [2.656] [2.022] [-0.314] [0.663] Observations 320 306 242 320 306 242 Number of country_id 42 38 40 42 38 40 R-squared 0.587 0.57 0.504 0.568 0.559 0.475 t statistics in brackets *** p<0.01, ** p<0.05, * p<0.1 Note: M.E. refers to the product between the sample mean and the coefficient for each variable. For interactive variables the M.E. is calculated only for countries with GDP per capital (PPP) threshold dummy equal to 1. Source: Authors’ calculation. 27 ME 3.2988 -0.0519 -0.0681 0.4136 0.3232 -0.2448 -0.5341 -1.0925 0.1995 -2.7539 0.9319 -0.6609 0.3104 Regressions in Table 7 allow isolating the effect that remittances can have for those countries where they are more important. With the 3.5 per cent threshold, the dummy variable is significant for two of the agencies. The interactive term for the volatility of external flows, also, is positive and significant. Increasing the threshold to 5 per cent does have an effect on the dummy variable and it does also affect the interactive variable, with a positive and significant effect on the sovereign rating. This result suggests that for countries where remittances are more important, high remittances have a direct effect on ratings (the dummy variable remittances over GDP is significant) but above all, and more relevant, they could have a premium in their ratings, meaning that there is an indirect effect of remittances on ratings (the interactive variable for the volatility of external flows is significant). Indeed, there is an insight. The indirect impact of remittances on ratings goes mainly through the volatility of external flows and not through the solvency ratio as argued in previous research. For countries where the ratio remittances-to-GDP is higher than 3.5 per cent the elasticity of the rating with respect to the variable volatility of external flows is 9 12 . Since 12 is positive, the weight of the variable volatility of external flows is reduced. We find that including an interaction term between the ratio remittances/GDP and the volatility of flows variable denotes a more inelastic rating for those countries where precisely remittances are more important. In other words, for countries in which the ratio of remittances over GDP is high there is an indirect effect of remittances. Indeed, there is a premium for these economies. More precisely, for countries in which remittances are relatively high, the negative impact of the volatility of external flows on the rating can be attenuated. In any case, as it is depicted in Figures 7a, 7b, and 7c, the effect is somehow limited. We estimate the predicted rating for Latin American countries, for the model including the threshold for remittances/GDP. The predicted ratings are presented in Figure 7a, 7b, and 7c. The main finding is that predicted ratings are higher than observed ratings for all speculative grade countries (except Bolivia). Source: Authors’ calculation. 29 Uruguay 12 6 0 12 10 6 4 2 Observed 0 Predicted 2007 8 2007 10 2006 Venezuela 2006 10 2005 12 10 9 8 7 6 5 4 3 2 1 0 2005 Panama 2004 2005 2006 2007 2005 2006 2007 2000 2004 0 2004 Predicted 2004 Observed 2 2003 6 2003 8 2002 10 2002 4 Predicted 2003 0 Mexico 2001 12 2001 Observed 2003 2 2002 4 2002 6 2001 14 2000 7 Observed 2000 0 1999 2 1999 Predicted 1999 4 2 1998 Observed 1998 14 1998 4 1997 8 1997 10 Predicted 2005 2004 2003 2002 2001 2000 1999 1998 1997 1996 1995 1994 1993 2007 2006 2005 2004 2003 2002 2001 2000 1999 2007 Observed 2006 6 2007 Costa Rica 2006 2005 2004 2003 2002 2001 2000 1999 1998 1997 1996 0 1997 Ecuador 1995 2 1996 0 1996 Predicted 1996 2 1994 4 Predicted 1995 Observed 1995 1993 4 1994 Observed 1995 Honduras 1993 8 Rating 10 8 1994 8 1993 6 Rating 2007 10 1994 5 Rating 2007 2006 2005 2004 2003 2002 6 1993 Rating 2007 2006 2005 2004 2003 2002 12 2001 8 Rating 2007 2006 2005 2004 2003 2002 2001 14 2000 Predicted Rating 12 1999 2 2001 Colombia 1998 Predicted 2001 12 1997 2 2000 2 1996 Observed 2000 1998 1997 4 Predicted 1995 4 2000 1999 1998 1996 Observed Observed 1994 Observed 1999 1998 1997 1995 1994 Bolivia 1993 4 2007 14 2006 0 2005 0 2006 2 2004 1 2005 0 Predicted 2003 4 2002 2 2004 8 Observed 6 2003 10 8 2002 5 2001 6 2001 14 2000 Nicaragua 2000 7.5 1999 1 1998 3 1999 9 1998 Predicted 1997 12 1999 Observed 1997 10 1997 9.5 1996 Guatemala 1996 0 1996 14 1998 Peru 1993 Rating 10 9 8 7 6 5 4 3 2 1 0 1997 6 1996 10 1995 Dominican Repubic 1995 12 1996 0 1995 4 1995 8 1994 12 1995 14 1994 Chile 1993 16 1994 2007 0 1993 10 Rating 2006 2005 2004 2003 2002 6 1993 8 Rating 2007 2006 2005 2004 2003 2001 2000 8 1994 9 Rating 2007 2006 2005 2004 2003 2002 2001 10 1993 Rating 2007 2006 2005 2004 2003 2002 2001 1999 Argentina 1994 6 Rating 2007 2006 2005 2004 2003 2002 2001 2000 1998 Rating 12 1993 2007 2006 2005 2004 2003 12 2002 Predicted 2002 3 2001 Observed 2001 4 2000 Predicted 2000 2 2000 Observed 2000 4 1999 18 1999 1997 Predicted 1999 1998 2 1999 1998 1996 Observed 1999 7 1998 1997 1995 1994 1993 4 1998 1997 1996 1995 1994 1993 Rating 14 1998 10.5 1997 1996 1995 1994 1993 Rating 2 1997 Rating 6 1997 8 1996 1995 1994 1993 8.5 1996 1995 1994 1993 Rating 14 1996 1995 1994 1993 Rating Figure 7a. Predicted vs. observed analysis for Latin America – S&P All Brazil 10 8 6 0 Predicted 12 El Salvador 10 8 6 Predicted 0 Paraguay Observed Predicted Source: Authors’ calculation. 9.5 Uruguay 14 12 10 10 8 6 8 0 Observed 4 Observed 6 4 Observed Predicted 2 Predicted 2 0 Predicted 2007 Venezuela 2007 1 2006 Predicted 2006 3 2005 Observed 2005 10 2004 10.5 2004 6 2003 7 2003 12 11.5 2002 6 2002 9 2001 Panama 2001 0 2006 2007 2006 2007 2007 2002 2001 2000 1999 1998 1997 2006 Predicted 2005 0 2005 Observed 2005 6 2004 8 2004 Mexico 2004 0 2003 Predicted 2003 6 2003 8 2002 10 2002 El Salvador 2001 2000 1999 12 2001 2000 2 2000 4 Predicted 2000 16 1999 Observed 1998 14 1999 3 1997 Ecuador 1999 10 1996 Colombia 1998 12 5 1995 2007 2006 2005 2004 2003 2002 2001 2000 1999 1998 1997 1996 1995 12 1998 6 1994 Rating Bolivia 1998 14 1997 Honduras 1997 0 1997 2 1996 4 Predicted 1996 Observed Observed 1996 4 12.2 12 11.8 11.6 11.4 11.2 11 10.8 10.6 10.4 10.2 10 1996 0 1994 10 1995 12 1995 0 1995 1993 2 1995 1993 Observed 4 Predicted 1994 6 1993 2007 2 Observed 1994 8 Rating 2006 2005 2004 4 1993 6 Rating 2007 2006 2005 2003 2002 2001 2000 6 1994 7 Rating 2007 2006 2005 2004 2003 2002 2001 1999 5 1993 11 Rating 2007 2006 2005 2004 2003 2002 2001 2000 7 1994 12 Rating 2007 2006 2005 2004 2003 2002 2001 2000 1998 1997 1996 8 1993 2007 14 2006 Peru 2005 0 2004 1 2004 2 2003 Predicted 2003 3 2002 7 2002 12.5 2001 0 2001 1 2000 2 2000 Predicted 2000 Observed 1999 Predicted 1999 Observed 2 1999 4 Predicted 1999 8 1998 Observed 1999 Guatemala 1998 12 1998 8 1997 14 1998 10 1995 Rating 9 1998 2 1997 4 1997 0 1996 1 1997 Predicted 1994 1993 3 1997 Observed 1996 8 1996 10 8 1996 10 1996 Dominican Repubic 1995 0 1995 4 1995 12 1995 14 1994 Chile 1995 Nicaragua 1993 16 1994 8 1993 2007 0 1994 10 Rating 2006 2005 2004 2003 2002 2001 2000 6 1993 6 Rating 2007 2006 2005 2004 2003 2002 2001 1999 10 1994 6 Rating 2007 2006 2005 2004 2003 2002 2001 2000 1998 8 1993 Observed Rating 2007 2006 2005 2004 2003 2002 2001 2000 1999 12 1994 Rating 2007 2006 2005 2004 2003 2002 2001 2000 1999 1997 Rating Argentina 1993 2007 2006 2005 2004 2003 2002 2001 2000 4 1999 18 2000 8 1998 Predicted 1999 5 1998 2 1999 12 1998 1996 1995 1994 1993 Observed 1998 1997 Rating 4 1998 10 9 8 7 6 5 4 3 2 1 0 1997 2 1997 2 1996 1995 1994 1993 6 1997 Rating 14 1997 4 1996 1995 1994 1993 2 1996 1995 1994 1993 Rating 4 1996 1995 1994 1993 Rating 12 1996 1995 1994 1993 Rating Figure 7b. Counterfactual Analysis for Latin America – Moody’s All 10 Brazil 8 6 0 Predicted Costa Rica Observed Predicted 8 Paraguay 5 4 2 Observed 0 Predicted Source: Authors’ calculation. 31 2 3 1 0 0 0 12 Uruguay 12 10 6 0 4 Observed 2 Predicted 2007 1 2006 4 2007 2 2005 10 2006 12 2004 9 2005 Panama 2003 2007 2006 2005 2004 Predicted 2004 0 2003 Observed 2 2002 4 Predicted 2003 Observed 2002 14 2001 Honduras 2002 2007 2006 2005 2004 2003 2002 2 2001 0 2001 Predicted 2000 2 2001 Predicted 2000 4 2 2000 Observed 1999 4 1999 Observed 1999 14 1998 Ecuador 1998 Observed 4 2007 2006 2005 2004 2003 2002 2001 2000 1999 1998 1997 12 1998 8 1996 Colombia 1997 6 1997 8 1997 10 8 1996 12 10 1996 9 1996 0 1995 9.5 1995 10 Predicted 1994 Observed 2 1995 4 Predicted 1995 6 1994 1993 Observed 1994 1993 8 Rating 10 1994 12 10 Rating 2007 2006 2005 2004 12 1993 6 Rating 2007 2006 2005 2004 2003 2002 2007 2006 2005 2004 2003 2002 2001 2000 1999 1998 1997 1996 1995 1994 1993 2007 2006 2005 2004 2003 2002 2001 0 1993 Rating 2007 2006 2005 2004 2003 2002 2 2000 8 Rating 2007 2006 2005 2004 2003 2003 2002 2000 Rating 14 1999 Predicted 2001 Bolivia 1998 2 2001 4 Predicted 1997 Predicted 2001 1999 1998 Observed 1996 Observed 2000 1999 1997 6 1995 4 2000 1999 1998 1996 8 1994 0 1997 10 1993 Observed 2007 8 2006 10 2005 14 2004 Peru 6 Predicted 2002 3 Observed 2003 Predicted 8 2001 6 2002 7 2000 14 2001 Nicaragua 2000 0 1999 2 2000 12 1999 Guatemala 1998 Predicted 1998 14 1999 4 1997 14 1998 Observed 1997 8 1995 Rating 12 1998 10 8 1996 2 1997 10 1996 0 1994 4 1997 Dominican Repubic 1996 0 1996 12 1995 14 1996 Chile 1995 16 1995 0 1995 Predicted 1994 1993 4 1994 Observed 1995 6 1993 2007 12 1994 8 1993 10 Rating 2006 2005 2004 2003 2002 2001 2000 14 1994 12 Rating 2007 2006 2005 2004 2003 2002 2001 1999 1998 8 1993 6 Rating 2007 2006 2005 2004 2003 2002 2001 2000 1999 1997 10 1993 Observed Rating 2007 2006 2005 2004 2003 2002 2001 2000 1999 1998 Argentina 1994 6 Rating 2007 2006 2005 2004 2003 2002 2001 2000 1999 1998 1996 1995 1994 1993 Rating 16 1993 2007 2006 2005 2004 2003 2 2002 4 2001 12 2000 4 2000 5 1999 8 1999 12 1998 18 1998 Rating 18 1998 2 1997 4 1997 1996 1995 1994 1993 6 1997 1996 1995 1994 1993 Rating 2 1997 Rating 6 1997 2 1996 1995 1994 1993 4 1996 1995 1994 1993 Rating 14 1996 1995 1994 1993 Rating Figure 7c. Predicted vs observed analysis for Latin America – Fitch All Brazil 12 10 8 6 Observed 0 Predicted Costa Rica 11.5 10.5 11 Observed Predicted El Salvador 6 0 Predicted Mexico 12 10 8 6 0 8 Paraguay 7 6 5 4 2 Observed Predicted 10 Venezuela 8 6 0 4. Conclusion This paper analyzes the impact of remittance flows on sovereign ratings for developing and emerging countries over the period 1993-2006. Remittances are an important source of external financing for developing countries. They may also have served to reducing country default risk most significantly in the smaller economies. In order to capture the impact of remittances on sovereign risk, we analyzed two core variables. First, we used a traditional solvency ratio employed in the literature and second we introduce a second determinant, the volatility of external flows. Using a model of the determinants of sovereign ratings and then a counterfactual estimation, we find that the impact of including remittances in both the ratio debt-to-exports and the volatility of external flows is weak. This suggests that CRAs are either not sensitive remittance flows, or that the inclusion of remittances in these two variables does not improve ratings. When specifying a second model for identifying highly remittance-dependent economies, we find that, for this set of countries, remittances have an indirect, yet positive impact on ratings through a premium (captured with the interactive dummy variable remittances over GDP and the volatility of external flows). The results presented above are consistent with previous research (e.g. Roubini and Manasse, 2005) showing that there is not a single model to rate countries and that not all variables have the same impact on a sovereign rating. For the remittances variable, the countries most affected are those where remittances account for a higher share of GDP. Indeed, the impact depends more on the volatility of external flows than on the solvency ratio (debt over exports). This paper also provides shadow ratings for countries which are not rated by the three main CRAs, in particular some Central American and Caribbean countries, where relative remittances flows are higher. Our analysis provides useful information on the potential ratings for some of these countries, a crucial step for accessing international capital markets. This approach implies that remittance flows cannot, as a whole, enable low and medium income countries to improve their creditworthiness significantly. However, remittances may prove very useful for some small Central American and Caribbean countries. Rating agencies, and other financial institutions, should take remittances into consideration to rate countries with idiosyncratic profiles. 5. References AFONSO A. (2003), “Understanding the Determinants of Sovereign Debt Ratings: Evidence for the Two Leading Agencies”, Journal of Finance and Economics, Vol. 27, No.1. ALTMAN E. (1989), “Measuring Corporate Bond Mortality and Performance”, Journal of Finance, Vol.XLIV, No.4. AMUEDO-DORANTES C. and S. POZO (2004), “Workers’ Remittances and the Real Exchange Rate: A Paradox of Gifts”, World Development Vol. 32, No. 8. CANTOR R. and C. MANN (2003), “Measuring the Performance of Corporate Bond Ratings”. CANTOR R. and F. PACKER (1994), “The Credit Rating Industry”, Federal Reserve Bank of New York, Quarterly Review, Summer/Fall, New York. CANTOR R. and F. PACKER (1996), “Determinants and Impact of Sovereign Credit Ratings”, Federal Reserve Bank of New York, Economic Policy Review, October, New York. CAVALLO E., A. POWELL and R. RIGOBON (2008), “Do Credit Rating Agencies Add Value? Evidence from the Sovereign Rating Business Institutions”, Inter-American Development Bank, Working Paper No.647, Washington, DC. CHAMI R., A. BARAJAS, T. COSIMANO, C. FULLENKAMP, M. GAPEN, and P. MONTIEL (2008), ”Macroeconomic Consequences of Remittances”, International Monetary Fund, IMF Occasional Paper No.259, Washington, DC. DUSTMANN C. and O. KIRCHAMP, “The Optimal Migration Duration and Activity Choice After ReMigration”, IZA Discussion Paper 299, Institute for the Study of Labor, Bonn, Germany. FAJNZYLBER P. and J. H. LOPEZ (2007), “Close to Home. The Development Impact of Remittances in Latin America”, World Bank, Washington, DC. FITCH RATINGS (2008a), “Complete Sovereign Rating History”, January 30. FITCH RATINGS (2008b), “Republic of El Salvador”, February 28. FITCH RATINGS (2009), fitchratings.com. GAILLARD N. (2009), “Fitch, Moody’s and S&P’s Sovereign Ratings and EMBI Global Spreads: Lessons from 1993-2007”, International Research Journal of Finance and Economics, April. GIULIANO P. and M. RUIZ-ARRANZ (2005), “Remittances, Financial Development, and Growth”, International Monetary Fund, IMF Working Paper 05/234. Washington, DC. JUTTNER J.D. and J. MC CARTHY (2000), “Modelling a Rating Crisis”, Macquarie University, Unpublished, Sydney. 33 KAMINSKY G. and S. SCHMUKLER (2001), “Emerging Markets Instability: Do Sovereign Ratings Affect Country Risk and Stock Returns?”, World Bank Working Paper 2678, Washington, DC. LEVICH R., G. MAJNONI and C. REINHART (2003), “Ratings, Rating Agencies and the Global Financial System”, Kluwer Academic Publishers. MASSEY D. and E. PARRADO (1998). “International Migration and Business Formation in Mexico.” Social Science Quarterly 79(1): 1–20. MATHIS, J., J MC ANDREWS, and J-C ROCHET (2008) “Rating the Raters”, Toulouse Working Paper. MOODY’S INVESTORS SERVICE (2004), “A Quantitative Model for Foreign Currency Government Bond Ratings”, February. MOODY’S INVESTORS SERVICE (2008), “Sovereign Default and Recovery Rates, 1983-2007”, March. MOODY’S INVESTORS SERVICE (2009), moodys.com. MORA N. (2006), “Sovereign credit ratings: guilty beyond reasonable doubt?”, Journal of Banking and Finance Vol.30. OSILI U. (2006), “Remittances and savings from international migration: Theory and evidence using a matched sample”, Journal of Development Economics, Vol 83, Issue 2, August. PORTES R. (2008), “Ratings Agency Reform”, VOX, 22. RATHA D. (2004), “Understanding the Importance of Remittance Flows.” Migration Policy Institute, Washington, DC. RATHA D., P. DE and S. MOHAPATRA (2007), “Shadow Sovereign Ratings For Unrated Developing Countries”, Development Prospects Group, World Bank, World Bank Policy Research Working Paper 4269, June, Washington, DC. RATHA D., S. MOHAPATRA and Z. XU (2008), “Outlook for Remittance Flows 2008‐2010: Growth expected to moderate significantly, but flows to remain resilient”, Migration and Remittances Team, Development Prospects Group, The World Bank, Migration and Development Brief No.8, November 11, Washington, DC. REISEN H. and J. VON MALTZAN (1999), “Boom and Bust and Sovereign Ratings”, OECD Development Centre, Working Paper, No. 148, Paris. ROUBINI N. and MANASSE P. (2005), ““Rules of Thumb” for Sovereign Debt Crises”, International Monetary Fund, IMF Working Paper 05/42, Washington, DC. ROWLAND P. and J. L. TORRES (2004), “Determinants of Spread and Creditworthiness for Emerging Market Sovereign Debt: A Panel Data Study”, Banco de la República de Colombia, Borradores de Economía, Bogota. STANDARD AND POOR’S (2005), “Sovereign Report Card: EMBIG”, Commentary Report, September 19. STANDARD AND POOR’S (2007), “Sovereign Ratings History Since 1975”, January 3. STANDARD AND POOR’S (2009), standardandpoors.com. SUTTON G. (2005), “Potentially endogenous borrowing and developing country sovereign credit ratings”, Bank for International Settlements, BIS Occasional Paper No.5, Basel. SY A. (2001), “Emerging Market Bond Spreads and Sovereign Credit Ratings: Reconciling Market Views with Economic Fundamentals”, International Monetary Fund, IMF Working Paper 01/165, Washington, DC. TOXOPEUS H. and R. LENSINK (2007), “Remittances and Financial Inclusion in Development”, UNUWIDER, Research Paper No. 2007/49, Groningen. WORLD BANK (2006), “Global Economic Prospects: Economic Implications of Remittances and Migration”. Washington, D.C. 35